©International Monetary Fund. Not for Redistribution€¦ · assistance from the Fund. The exchange systems of The Gambia, Guyana, Jamaica and Nigeria are currently classified by

IMF WORKING PAPER

© 1993 International Monetary Fund

This is a Woridng Paper and the author would welcome anycomments on the present text. Citations should refer to aWorking Paper of the International Monetary Fund, men-tioning the author, and the date of issuance. The viewsexpressed are those of the author and do not necessarilyrepresent those of the Fund.

WP/93/36 INTERNATIONAL MONETARY FUND

Policy Development and Review Department

Experience With Floating Interbank ExchangeRate Systems in Five Developing Economies

Prepared by Vicente Galbis I/

April 1993

Abstract

This paper reviews the experience with floating interbank exchange ratesystems in five developing countries--The Gambia, Guyana, Jamaica, Nigeriaand Sri Lanka--and draws some conclusions about the stability and efficiencyof these systems. The experience of these countries illustrates both thedifficulties and the advantages of interbank exchange rate markets. Themain conclusion is that these markets can operate relatively well with aminimum banking infrastructure, provided that the authorities remove legaland institutional impediments to the free operation of these marketsincluding, in particular, exchange restrictions. Any residual restrictionsthat may remain will likely give rise to the continued existence of parallelmarkets.

JEL Classification Numbers:

F31, 016, 057

I/ I would like to thank Anupam Basu, Evangelos A. Calamitsis, MariaCarkovic, J.R. Dodsworth, Hans-Michael Flickenschild, Manuel Guitian,E. van der Mensbrugghe, Marjorie B. Rose, Mohammad Shadman, andBrian C. Stuart for their valuable comments and additional information.However, the responsibility for any remaining errors rests solely with theauthor.

©International Monetary Fund. Not for Redistribution

- ii -

I.

II.

Ill,

IV.

V.

Tables

1.2.3.4.

Charts

1.2.3.4.5.

Contents

Summary

Introduction

Initial Conditions and Introduction of the Floats

1. Exchange rate systems and economic conditionsbefore the float

2. Factors affecting the choice of an interbankfloating rate system

3. The role of financial institutions and theregulatory framework

Reforms of the Exchange and Trade System

1. Macroeconomic adjustment framework2. Liberalization of the exchange system3. Reform of the trade regime

Performance of the Interbank Floats

1. Exchange rate developments2. Developments in the external sector and the

general economy

Conclusions

Changes in the Official Exchange RateParallel Exchange Market DevelopmentsExternal IndicatorsGeneral Economic Indicators

The Gambia: Exchange RateGuyana: Exchange RateJamaica: Exchange RateNigeria: Exchange RateSri Lanka: Exchange Rate

Page

iii

1

2

2

9

10

12

121419

20

21

24

25

22a22b22c22d22e

3567


- Ill -

Summary

This paper examines the experiences under floating interbank exchangerate systems of five developing countries--The Gambia, Guyana, Jamaica,Nigeria, and Sri Lanka--which employed a variety of institutional andregulatory arrangements in the interbank market. It describes theirprefloat economic conditions and exchange arrangements and the factorsaffecting their choice of an interbank system. The role of financialinstitutions and the official regulatory framework in making the transitionto the interbank exchange rate system is also discussed, along with themacroeconomic adjustment framework and the exchange and trade system reformsadopted. The paper assesses the performance of the interbank floats interms of both exchange market and macroeconomic developments.

All five countries introduced an interbank floating exchange system toprovide a more efficient, noninterventionist mechanism for determining theofficial rate and allocating scarce foreign exchange resources. However,some phased in the new system, removed exchange and trade restrictions, andliberalized interest rates more gradually than did others; in severalcountries, too, remaining regulations constrained banks and other authorizedexchange dealers in negotiating the exchange rate between themselves and thepublic. The interbank floats and the accompanying exchange and tradeliberalizations generally resulted in convergence between the official andparallel exchange rates; an appropriate flexibility and movement of nominaland real exchange rates; and spreads between the buying and selling ratesthat stayed within reasonable limits. Economic gains from the floatinginterbank systems also appear to have accrued more visibly to the countriesthat moved early and decisively to abandon previous restrictions--TheGambia, Guyana, Jamaica, and Sri Lanka.

The paper concludes that interbank exchange rate markets can operaterelatively well with minimal banking infrastructure (for example, TheGambia and Guyana); licensing of nonbank foreign exchange dealers canprovide additional market competition. However, the paper argues, thesemarkets can operate well only if the authorities remove trade barriers andexchange restrictions on both current international transactions and capitaltransfers, liberalize interest rates, and introduce prudential regulationsand supervision over exchange transactions. Segmented exchange markets maypersist if regulations prevent the free flow of resources across marketparticipants (as shown by the experience of Nigeria with a compositeexchange rate system--an official auction and an interbank market).Moreover, any residual official restrictions that remain are likely tofoster the continued existence of parallel, informal markets. Finally, thepaper concludes, floating exchange rates and liberalized exchange and tradesystems--although they are helpful in balancing the external sector--are notsubstitutes for the sound macroeconomic policies.


This page intentionally left blank


I. Introduction

In the 1980s some developing countries began to introduce their ownmarket-determined floating exchange rate systems. Two forms of floatingarrangements were adopted, the interbank and auction systems, i/ Themajority of the countries that introduced a floating exchange rate opted forthe interbank system consisting of an exchange market operated by commercialbanks and licensed foreign exchange dealers. Other countries, largely dueto different institutional constraints, adopted an official auction systemunder which foreign exchange was surrendered to the central bank, which thenauctioned the exchange to the commercial banks and/or the nonfinancialsector. This paper examines the experience under floating interbankexchange rate systems in five countries--The Gambia, Guyana, Jamaica,Nigeria and Sri Lanka- -which were chosen maitvly because of the variety oftheir experience in the setting up of the institutional and regulatoryframework in the interbank market. 2/

The paper is organized as follows. Section II describes the initialeconomic conditions and the exchange arrangements of the five countriesprior to the adoption of the floating exchange rate systems and the factorsaffecting the choice of an interbank system. The role of both financialinstitutions and the official regulatory framework in making the transitionto the interbank exchange systems is also discussed. Section III reviewsthe macroeconomic adjustment framework and the exchange and trade systemreforms adopted in the five countries. The performance of the interbankfloats in terms of both exchange market and macroeconomic developments isexamined in Section IV. The main conclusions are contained in the finalsection.

i/ For a general overview of these systems, see Peter J. Quirk, BenedicteVibe Christensen, Kyung-Mo Huh, and Toshihiko Sasaki, Floating ExchangeRates in Developing Countries - Experience with Auction and InterbankMarkets. IMF Occasional Paper No. 53 (May 1987).

2J In all five cases the authorities availed themselves of technicalassistance from the Fund. The exchange systems of The Gambia, Guyana,Jamaica and Nigeria are currently classified by the Fund as independentlyfloating, whereas that of Sri Lanka is classified as other managed floating.Many other developing countries have adopted floating exchange rates inrecent years.


- 2 -

II. Initial Conditions and Introduction of the Floats

1. Exchange rate systems and economic conditions before the float

Prior to the adoption of floating interbank exchange rate systems, eachof the five countries surveyed had been operating under a pegged exchangerate arrangement. The currencies of Guyana, Jamaica, Nigeria and Sri Lankawere pegged to the U.S. dollar, whereas the Gambian dalasi was pegged to theUK pound sterling. In the pre-float period, the official exchange rates ofthe countries pegged to the dollar were depreciated on several occasions;and The Gambia's authorities accepted the nominal effective depreciationthat was taking place because of the tendency of the U.K. pound to fall inworld markets (Table 1). \J In deciding to depreciate the rate, there areindications that the authorities were trying to address a variety ofimmediate concerns or objectives, including protecting externalcompetitiveness in the face of substantial domestic inflation; narrowing thedifferential between the official and the parallel market rates to preventthe diversion of foreign exchange to the parallel market; and reversing theprolonged deterioration in the foreign reserve position. Nevertheless,owing to the prevailing exchange and trade restrictions and expansionaryfinancial policies, these objectives were generally not achieved.

In all cases, in response to the scarcity of foreign exchange in theofficial market, the authorities had tightened exchange and traderestrictions. Surrender requirements were applied to all exports, andcomprehensive restrictions were imposed on capital outflows. In addition,there were exchange restrictions on current transactions (typically forservices payments); and official foreign exchange resources were allocatedfor imports through cumbersome licensing and other nonprice administrativeprocedures. In The Gambia, for instance, the trade system includedcomprehensive import licensing, with specific import restrictions on productgroups and countries, as well as tariffs that were high and discriminatory.In Jamaica there was an informal rationing of documented imports and afairly comprehensive system of exchange controls on current payments andcapital transfers. Similarly, in Guyana and Nigeria, the trade system -included import prohibitions and a discriminatory import licensing system.

In most cases, such restrictions had various deleterious effects,including, inter alia, the diversion of transactions to the parallel market;a persistent weakening of export performance and the overall balance of

\J While in all cases the authorities were particularly concerned aboutthe political repercussions of depreciation, in the case of The Gambia,given neighboring Senegal's fixed link to the French franc and theimportance of border trade, the authorities also felt that a fixed peg to amajor convertible currency was crucial to maintaining confidence in thedomestic currency and to avoiding uncertainty and disruptions in externaltransactions.


Table 1. Changes in the Official Exchange Rate

(In local currency per U.S. dollar)

CurrencyDate of introduction peg

of float before

Rate of exchange rate depreciation (+) or appreciation (-)(Based on end-year exchange ratesin local currency per U.S. dollar)

1.

2.

3.

4a

4b

5.

Country

The Gambia

Guyana

Jamaica

Nigeria

Nigeria

Sri Lanka

(t year)

Jan.

March

Sept.

Sept.

Jan.

Aug.

20,

13

17

26

9,

20,

1986

, 1990

, 1990

, 1986

1989

1990

float

UK£ I/

US$

US$

US$

n.a. 2/

US$

t-

11

127

0

11

231

7

3

.3

.3

.4

.7

.7

.9

t-2 t-1(Percentage change)

56.8 80

230

-0.4 18

7.9 23

24.8 29

7.4 21

.1

.0

.2

.8

.3

.1

t

114

36

24

231

42

0

.6

.4

.0

.7

.0

.6

Sources: IMF, International Financial Statistics: and staff documents.

!/ The depreciation of the Gambian dalasi with respect to the U.S. dollar in the years 1983-85 reflectsthe depreciation of the U.K. pound with respect to the dollar. During this period the rate of the dalasiwith respect to the U.K. pound was not altered.

2/ The system was changed from one float to another, as described in the appendix.


- 4 -

payments; shortages of inputs and underutilization of capacity in import-dependent productive sectors; and the worsening of inflationary pressuresand relative.price distortions. For example, the exchange markets in thefive countries were characterized de facto by a high degree of segmentationand differential exchange rates (Table 2). I/ In The Gambia, it wasdifficult for the authorities to prevent the increasing diversion of foreignexchange resources into the parallel market. In Guyana, substantialinflationary pressures and a large premium in the parallel exchange ratemarket penalized exports and led to significant relative price distortionsbetween different import uses. Despite the tight restrictions on currentand capital transactions, there was substantial capital flight out ofJamaica. Nigeria's official exchange rate for the naira became increasinglyovervalued on account of rapid domestic inflation, and the balance ofpayments came under severe pressure. In the case of Sri Lanka, too, theauthorities were faced with prolonged balance of payments pressures. In allcases, the move to the floating system was preceded by adverse developmentsin the current account, and a virtual exhaustion of official foreignexchange reserves (Table 3). 2/

In all the sample countries, the pre-float period was marked byexpansionary financial policies. This was reflected in the high ratios offiscal deficits to GDP, and in the rapid growth rates of domestic credit(particularly to the public sector) and of broad money. Real interest rateswere generally negative (except in Jamaica) because of financial repressionpolicies, which provided incentives for capital outflows. At the same time,the use of nonmarket, direct instruments of monetary policy produceddistortions in the structure of interest rates, and in the allocation ofcredit, which adversely affected the competitiveness of the banking system,and led to weak savings and unproductive investments.

The data show that in the pre-float period, there was both substantialinflation and a generally weak real growth performance (Table 4). Realexport growth was generally faltering and, although data are only sketchy,there was also a tendency toward limited growth in import volume, a factorthat probably contributed to the weak growth performance.

I/ Table 2 shows the evolution of the main official and parallel marketrates and of the premium in the parallel market. It is based on exchangerates for,the U.S. dollar as reported in IPS for the main official rate andas reported by International Currency Analysis, Inc. for the street exchangerate. In some cases, e.g., The Gambia, the street parallel market rate isnot an appropriate indicator of the market behavior. Also, it should benoted that in some countries, e.g., Nigeria, for some periods, the existenceof several legal exchange rates besides the parallel rate complicates thepicture substantially. Unfortunately, it is difficult to obtain data forall the various rates. For this reason the data should be interpreted withcaution.

2/ Accumulation of arrears on external debt payments was also pervasive.


- 5 -

Table 2. Parallel Exchange Market Developments

(Local Currency Units per U.S. Dollar1

CountryPre-float periodt-3 t-2 t-1

Year offloat

Post-float periodt+1 t+2 t+3

1. The Gambia

Official rateParallel ratePremium in parallel market \J

2. Guyana

Official rateParallel ratePremium in parallel market 2/

3. Jamaica

Official rateParallel ratePremium in parallel market

4a. Nigeria


4b. Nigeria

Official rateParallel ratePremium in parallel market 5_/

5. Sri Lanka

2.757 4.323

10.351.0395.1

5.5006.50018.2

10.053.0

430.0

5.4806.70022.3

0.749 0.808

3.4613.7047.0

33.062.053.2

6.4808.30028.1

1.0003.704270.4

3.3177 .692131.9

4.141 5.3535.000 10.00020.7 86.8

(1286)

7.4268.33312.2

(1990)

45.0105.0133.3

(1990)

8.03810.200

26 .9

(1986)

3.3177.692131.9

(1989)

7.6519.09118.8

(1990)

6.4268.33329.4

122.0168.037.7

21.49323.000

7.0

4.1415.00020.7

9.00111.11023.4

6.659 8.3159.091 7.14336.5 -14.1

126.0 3/175.0 3/38.9 3/

22.173 4/24.000 4/

8.2 4/

5.353 7.65110.000 9:09186.8 18.8

9.86212.50026.7

19.380 6/28.10 6/45.0 6/


30.76331.250.4

33.03345.0036.2

40.00048.0020.0

40 . 24050.2400.3

42.58043.100.0

44.63052.1600.5

!/ ...3/I/Z/...

Sources: IMF, International Financial Statistics: 1988-1989 World Currency Yearbook, byPhilip P. Cowitt; and Free or Black Market Rates of Payments and/or Transfers Abroad.monthly sheet, International Currency Analysis, Inc. (New York).

I/ An alternative series, estimated by the Central Bank based on the Central Bank'sfixing rate versus the parallel rate both in dalasis per pound sterling, is as follows:70.0; 0.9; 2.9; 2.5; 1.8. This alternative series appears to be more representative of themarket than the one above.2/ Alternative information from the Bank of Guyana suggests that the parallel market rate

typically does not currently differ significantly from the cambio market rate3/ November 1992.4/ August 1992.5_/ An alternative series based on the central bank rate versus the market bureau de

change exchange rate is as follows: 1990: 13.9; 1991: 58.6; September 1992- 19 56/ October 1992.7_/ The premium was 1.9 percent as late as August 1992.

t


- 6 -

Table 3. External Indicators

1.

2.

3.

4a

4b

5.

Country

The GambiaCurrent account deficit (+) Inrelation to exports (percent)

Export* In relation to reserves(ratio)

Reserves In relation to Imports(percent)

Balance of payments deficit (+) Inrelation to reserves (ratio)

Guyana I/Current account deficit (+) Inrelation to exports (percent)

Exports In relation to reserves(ratio)


Balance of payments deficit (+) Inrelation to reserves (ratio)

JamaicaCurrent account deficit (+) Inrelation to exports (percent)

Exports In relation to reserves(ratio)


Balance of payments deficit (+)In relation to reserves (ratio)

. NigeriaCurrent account deficit (+) Inrelation to exports (percent)

Exports in relation to reserves(ratio)

Reserves in relation to imports(percent)

Balance of payments deficit ( + )in relation to reserves (ratio)

. NigeriaCurrent account deficit (+) inrelation to exports (percent)

Exports in relation to reserves(ratio)

Reserves in relation to imports(percent)


Sri LankaCurrent account deficit ( + ) in

relation to exports (percent)Exports in relation to reserves(ratio)

Reserves In relation to imports(percent)


Source: IMF. International Financial

frnt-3

60.

18.

3.

7.

45.

28.

3,

16

19

4

16

-1

42

10

8

3

-6

5

29

0

23

5

14

0

-float

4

7

3

6

2

,6

.2

.0

.3

.1

.4

.7

.1

.5

.6

.0

.1

.6

.2

.9

.4

.0

.9

.2

t-2

-8.

40.

2.

8.

43.

53.

1.

26

-3

6

11

-0

-1

8

16

0

1

6

28

3

26

6

11

0

period

9

1

3

8

7

,2

.9

.2

.4

.0

.9

.5

.0

.1

.5

.7

.0

.5

.4

.9

.7

.7

.0

.5

t-1

-11.

36.

2.

3.

55.

15.

6,

13,

29

9

6

1

-19

7

22

0

2

10

15

7

27

6

11

-0

Year offloat Post-float period

6

3

8

3

1

.4

.3

.6

.7

.3

.7

.8

.6

.9

.2

.7

.8

.6

.2

.6

.5

.2

.9

.2

t t+1 t+2

(12M)

-6.5 -8.2 -31.9

4.8 2.9 4.4

16.0 27.1 18.0

2.3 -0.4 -0.9

(liiG)

72.5 50. OP

7.1 1.9P

11.5 50. 8P

6.8 0.5P

(1120)

22.8

6.9

10.4

-0.6

(1986)

-6.1 1.0 2.8

5.6 6.5 10.6

29.2 28.4 15.2

0.9 3.9 7.6

(1989)

-13.9 -37.7

4.5 3.5

47.8 78.3

1.5 -0.2

(liiS)

16.0

4.4

18.2

-0.1

t+3

-15.0

4.9

16.4

-0.1

-13.9

4.5

47.8

1.5

Statistics .

I/ IMF staff estimates, except for reserves taken from 1££. P - preliminary.


- 7 -

Table 4. General Economic Indicators

CountryPre-float periodt-3 t-2 t-1

Year offloat

tPost-float periodt+1 t+2 t+3

The Gambia

Real GDP growth rate I/ 2_/Export growth rate (volume) 2/ 3_/Real import growth rate 2/ 3_/Inflation rate (CPI based)Fiscal deficit (+)/GDP I/ 2/Money growth rate (based on money

plus quasi money)Increase in domestic credit aspercent of broad money

Increase in credit to public sectoras percent of broad money

Deposit interest rate

Guyana

Real GDP growth rate 3_/Export growth rate (volume) 3/Real import growth rate 3/Inflation rate (CPI based) I/Fiscal deficit (+)/GDPMoney growth rate (based on money




Jamaica

Real GDP growth rateExport growth rate (volume)Real import growth rate 3_/ 4/Inflation rate (CPI based)Fiscal deficit (+)/GDP 3/ A/Money growth rate (based on money




(1986)

-8.2-6.027.410.56.0

26.7

61.3

33.98.5

1-34-9227

5

39

269

.6

.7

.2

.1

.5

.5

.2

.4

.0

49-4184

51

67

359

.1

.0

.1

.3

.7

.3

.4

.4

.8

2-59

565

7

-61

-5316

.8

.3

.7

.6

.1

.3.

.7

.9

.1

1450237

24

-66

-6315

.7

.5

.3

.5

.2

.6

.3

.7

.8

4.24.18.11.-1.

14.

10.

5.15.

35078

7

5

40

5.221.519.28.31.7

20.8

-6.4

-13.012.9

(1990)

0.-3.11,28,40,

53.

79.

67,11,

,1.3.1.7.0

,8

,8

.2

.1

-3-6

-133929

41

29

1412

.0

.0

.3

.9

.8

.5

.7

.2

.0

-4-0-5

89.8

48

-20

-38t

.8

.9

.37E.1

.5

.9

.9

-6-82

63.35

84

20

029

.2

.8

.86E.4.

.1

.7

.1

.2

6.17.

t

105.

•

76

9

-11

IP5P

9E

.5

.6

.4

(1990)

6,6,14.6,

12,

11,

25,17,

.2

.9

.4

.6

.5

.8

.3

.5

1,-0,20,8,12

32

2

-19,17,

.5

.5

.6

.2

.8

.3

.6

.3

.9

429,12146

6

6

-1119

.6

.1

.4

.4

.3

.6

.1

.4

.0

318-2212

21

-3

-1626

.8

.1

.6

.9

.9

.5

.6

.0

.0

3/ .3-4510

51

0

-26. 27

.5

.0

.1

.4

.4

.4

.0

.4


Table 4 (Concluded).

- 8 -

General Economic Indicators

Pre-float periodCountry

4a. Nigeria

Real GDP growth rateExport growth rate (volume) (crudepetroleum)

Real import growth rateInflation rate (CPI based)Fiscal deficit (+)/GDPMoney growth rate (based on moneyplus quasi money)

Increase in domestic credit aspercent of broad money



4b. Nigeria

Real GDP growth rateExport growth rate (volume) (crudepetroleum)

Real import growth rateInflation rate (CPI based)Fiscal deficit (+)/GDPMoney growth rate (based on money




5. Sri Lanka

Real GDP growth rateExport growth rate (volume)Real import growth rateInflation rate (CPI based)Fiscal deficit (+)/GDPMoney growth rate (based on money




Source: IMF. International FinancialI/ Data from IMF Occasional Paper No.

t-3

-5.4

-7.4. . .

23.3. . .

14.0

34.6

28.47.4

3.1

-38.0

5.73.4

1.9

17.4

2.39.2

1.51.52.47.78.7

15.4

21.7

16.011.5

Statistics. 100.

t-2

-5.1

17.3. . .

39.64.6

11.6

15.7

13.28.3

-0.5

38.7

11.48.9

22.4

16.2

9.913.1

2.7-2.9-4.914.012.7

14.9

38.8

24.113.2

t-1

9.4

13.6. . .7.42.8

9.0

7.4

2.79.1

9.9

11.6

54.5

32.9

37.5

21.9

2.32.0-6.211.57.4

13.3

14.0

10.716.4

, except as

Year offloatt

(1986)

3.1

-38.0• . .5.73.4

1.9

17.4

2.39:2

(1989)

5.3

50.4

10.7

-18.5

-24.1

(1990)

6.216.55.621.57.8

15.4

24.5

8.019.4

Post-floatt+1 t+2

-0.5 9.9

38.7 11.6... ...

11.4 54.58.9

22.4 32.9

16.2 37.5

9.9 21.913.1

8.2 5.0

... ...

7.4 13.0

35.4

42.8

30.514.9

4.84.213.712.27.5

18.5... ...

periodt+3

5.3

. . .50.4

10.7

-18.5

-24.1

3/ ...

. . .indicated otherwise.

2./ Based on years from July to June. Example: 1986 refers to year July 1986-June 1987over preceding period.

3_/ IMF staff estimates. P - preliminary; E - estimate.4/ Information available on a fiscal year basis (April-March).


- 9 -

2. Factors affecting the choice of an interbank floating rate system

In deciding to adopt a floating exchange arrangement in The Gambia,Guyana, Jamaica, and Nigeria, the national authorities were concernedprimarily about eliminating the segmented nature of the exchange system.Moreover, the past policy of periodic depreciations under a fixed-peg regimehad not helped to solve the persistent balance of payments problems. In allfive countries, the official foreign reserve position had been so seriouslyweakened and the need for structural reforms in the foreign exchange andtrade regime was so pressing that it would have been difficult for therespective central banks to rely on a fixed exchange rate system. I/Given the magnitude of the desired correction in the external imbalances andthe need to substantially liberalize the exchange and trade systems, theunderlying path of the equilibrium exchange rate would have been difficultto forecast. In these circumstances, a market-based floating exchange ratesystem was expected to provide a more efficient and reliable mechanism fordetermining the official rate and allocating scarce foreign exchangeresources.

Each of the countries surveyed could, in principle, have chosen betweentwo types of floating exchange rate systems, namely, an interbank or anauction market. Apart from institutional considerations, the interbanksystem was chosen because it was more consistent with a market approach,minimized the scope for administrative interference, and permitted theauthorities to avoid the adverse political repercussions of announcingchanges in the exchange rate. However, a pure floating interbank exchangearrangement was initially introduced only in The Gambia (January 20, 1986)and in Jamaica (September 17, 1990). The other countries surveyed adopted amore phased approach to introducing a fully floating exchange rate system,mainly to fit their specific institutional settings, to apply an appreciatedrate for selected transactions, and to minimize market instability. Thus,in Guyana starting on March 13, 1990, as a transitional arrangement, afloating interbank exchange market covered all but certain specifiedtransactions, for which there was an officially managed fixed exchange rateuntil its elimination on February 21, 1991; thereafter, the exchange ratewas allowed to float freely. In Nigeria, starting on September 26, 1986, atwo-tier system was introduced as a transitional arrangement, with the firsttier comprising an administered exchange rate (for oil exports and certainpublic sector transactions), and a second-tier foreign exchange market(SFEM)--a composite market--comprising both a central bank auction forcommercial banks and a floating interbank exchange market. This system wasnot fully unified until March 1992. In Sri Lanka, starting on August 20,1990, financial institutions in the interbank exchange market were allowed

\J For example, the Sri Lankan authorities were interested in choosing anexchange rate regime that could both facilitate timely adjustments in theofficial exchange rate (by depoliticizing such actions) and better promotethe structural reform process; the parallel exchange market per se was notthe focus of their attention.


- 10 -

freely to negotiate rates with their customers within a 3 percent band oneither side of a reference rate, which was determined daily by the centralbank on the basis of a transaction-weighted average of the exchange ratesreported during the previous working day; free floating was thuscircumscribed within a fixed but adjustable band in order to avoid excessiveinstability in the rate. The potential scope for central bank management ofthe auction rate in Nigeria and of the reference rate for commercial banksin Sri Lanka, made these hybrid systems more prone to being officiallyadministered than (and hence, different from) the free floating interbankexchange markets of The Gambia and Jamaica.

3. The role of financial institutions and the regulatory framework

At the time of the introduction of their respective interbank exchangerate systems, the institutional environment of the financial sector was lessdeveloped in The Gambia and Guyana than in the other three countries. Therewere basically only three commercial banks in The Gambia, and five inGuyana, whereas Jamaica, Nigeria, and Sri Lanka had a larger number ofcommercial banks and also had other significant financial institutions,including merchant banks. Thus, in order to increase the number ofparticipants in the market in Guyana and foster competition among them, thedecision was made from the start to license 23 nonbank foreign exchangedealers (in addition to the 5 commercial banks). I/ In The Gambia,although only the three commercial banks were initially allowed to operatein the interbank market, starting in April 1990 foreign exchange bureausbegan to be licensed. In Jamaica, all 11 commercial banks, 19 merchantbanks, and a number of nonbank financial institutions, were authorized tooperate in the interbank exchange market from the start. In Nigeria, theoriginal interbank market (1986) was open only to commercial banks andmerchant banks, but by 1989 the reformed interbank market was also open toforeign exchange bureaus. In Sri Lanka many other financial institutionsbesides commercial banks, were allowed to operate in the market.

The central banks also assumed the role of main regulators of theinterbank exchange markets with a view to promoting their development,maintaining market competition, and providing adequate prudential regulationand supervision. To promote the interbank market, the authorities of thesecountries generally required that all transactions be undertaken bycommercial banks and other authorized dealers, with the exceptions being thetransactions of the government and/or of other specific sectors that were

\J The introduction of the interbank market was facilitated by theprevious existence of a well-developed parallel (cambio) market. The marketmechanisms, therefore, already were largely in place and licenses weregranted to nonbank foreign exchange dealers who previously had been activein the cambio market.


- 11 -

handled by the central bank (in most countries). I/ In The Gambia, topromote the interbank market, the authorities initially further requiredthat public enterprises surrender their foreign exchange earnings to thebanks; began to plan the government's purchases of foreign exchange from theCentral Bank within a framework of a foreign currency budget to enable theCentral Bank to smoothly phase its purchases in the interbank market;requested hotels involved in the tourist trade to repatriate their exchangeearnings through the banking system; and recognized the critical role ofappropriately supporting macroeconomic policies (including positive realinterest rates) in preventing capital outflows.

Prudential regulations and market supervision were developed todifferent degrees in the five countries. To protect the banks, theauthorities of The Gambia, Jamaica and Sri Lanka introduced guidelinesplacing limits on their foreign exchange working balances and openpositions. In Jamaica, dealers were required to post their rates to thepublic and to report the rates to the Bank of Jamaica; the Bank alsoprovided guidelines to prevent market abuse and ensure the liquidity andsolvency of dealers (by prescribing limits on their operations in relationto their capital). In Guyana, central bank supervision to date has beenlimited to the requirement that foreign exchange dealers report weekly onthe volume of transactions, the average exchange rate, and the opening andclosing balances.

Interest rate policies were liberalized prior to the introduction ofthe interbank exchange rate system in The Gambia (September 1985), Jamaica(1984) and Guyana (April 1989) and shortly after the introduction of the newexchange system in Nigeria (1987); substantial progress toward thisobjective is currently being made in Sri Lanka. The authorities also beganto develop appropriate free-market financial instruments to deepen themarkets and increase the flexibility of the rates. 2/ The freeing ofdomestic interest rates or the more active management of these rates quicklyincreased the rates toward positive real levels in all cases but Sri Lanka.The increase in domestic interest rates reduced incentives for capitaloutflows and, therefore, facilitated the progressive dismantling ofregulations designed to prevent these outflows, some of which could haveadversely affected the determination of the interbank exchange rate and itsstability.

I/ In Sri Lanka, the authorities permitted some domestic entities (e.g.,the Greater Colombo Economic Council and other approved enterprises) toretain the proceeds of their foreign borrowing in special accounts held inForeign Currency Banking Units, a measure taken with a view to promotingforeign capital mobilization.2y Mention may be made, for instance, of the introduction of Treasury

bill auctions in The Gambia in July 1986 and in Guyana in June 1991.


- 12 -

III. Reforms of the Exchange and Trade System

1. Macroeconomic adjustment framework

The floating interbank exchange rate systems under review wereintroduced in the context of medium-term programs of macroeconomicadjustment and structural reform. In all cases, emphasis was put on priceliberalization; opening up the trade and exchange system; restructuring andreducing the role of the public sector; and policies to promote thedevelopment of the financial sector and the nonfinancial private sector.Appropriately tight fiscal and monetary policies were programmed to providean environment conducive to durable structural reform. The broad objectivesof the programs were to provide the basis for sustained economic growth inthe context of low inflation and a viable balance of payments position overthe medium term. In each of the countries surveyed, the introduction of afloating interbank exchange market was combined with a significantliberalization of the exchange and trade system, because it was a necessarycondition for developing a free exchange market (including eventually aforward exchange market). I/ Equally important, it provided the basis foraligning domestic prices of tradeable goods with world prices. However,implementation of the programs varied from country to country and thephasing of the reforms was more gradual in some cases than in others.

In The Gambia, the authorities in June 1985 began to implement acomprehensive adjustment program, the Economic Recovery Program (ERP), aimedat restoring internal and external equilibrium and lay the foundations forsustainable economic growth. 2/ This was followed by the Program forSustained Development in 1990. These adjustment efforts were supported bysuccessive arrangements from the Fund beginning with two annual arrangementsunder the structural adjustment facility (SAF) during 1986/87-1987/88(July/June). The reforms helped to restore an appropriate structure ofrelative prices; introduce complementary structural reforms; and weresupported by restrictive fiscal and monetary policies.

Similarly, in Guyana, an Economic Recovery Program (ERP) was introducedin mid-1988 for a three-year period. As implementation of this initialprogram was weak, the government strengthened its reform efforts in thecontext of an updated ERP for 1990-92, prepared in collaboration with thestaffs of the Fund and the World Bank. Under the ERP the overall fiscaldeficit was reduced; the exchange and trade system was liberalized and the

I/ A brief discussion of forward markets is provided in the next section.For a comprehensive review of efforts to develop forward markets indeveloping countries, see Peter J. Quirk, Graham Hacche, Viktor Schoofs, andLothar Weniger, Policies for Developing Forward Foreign Exchange Markets.IMF Occasional Paper No. 60 (June 1988).2/ Michael T. Hadjimichael, Thomas Rumbaugh, and Eric Verreydt, The

Gambia: Economic Adjustment in a Small Open Economy. IMF Occasional PaperNo. 100 (October 1992).


- 13 -

exchange rate unified; virtually all price controls were removed andconsumer subsidies eliminated; the incentive framework for the privatesector was improved and state intervention was curtailed; progress was madein restructuring and reducing the size of the public sector; a shift tomarket-based instruments of monetary control was initiated; and arrears tomultilateral and bilateral donors were cleared.

Beginning in 1989, Jamaica implemented stabilization measures and wide-ranging structural reforms in the context of successive stand-byarrangements with the Fund, culminating in an extended arrangement for athree-year period beginning in October 1992. The broad objectives of theseprograms were achieved by further opening up the economy and by improvingmarket incentives through decontrolling prices, the exchange rate andinterest rates.

In early 1986, the Nigerian authorities drew up a medium-termstructural adjustment program (SAP) and in November 1986 they followed thisup by requesting a 14-month stand-by arrangement from the Fund. A majorcomponent of the SAP was the establishment of a new exchange rate system andthe liberalization of the exchange and trade system. Capital restrictionswere to be liberalized over the long run. The SAP removed price controls,liberalized the export of non-oil commodities, adjusted the prices of manyproducts, and terminated agricultural commodity boards. I/ However, inthe context of several successive arrangements with the Fund, there wereintermittent periods of excessively expansionary financial policies and theliberalization of the external sector was phased very gradually.

In 1988, Sri Lanka began to develop a medium-term strategy aimed ataddressing the structural constraints to higher growth and reducingmacroeconomic imbalances. This was supported by a three-year arrangementunder the Fund's structural adjustment facility (SAF) and funds frommultilateral and bilateral sources. After a disappointing start in thefirst year of the adjustment program, the Government recommenced itsstabilization and adjustment efforts in mid-1989. A pronounced tighteningof both fiscal and monetary policies was effected, and structural reformsbegan to be implemented including a change in the exchange rate regime andthe liberalization of the exchange and trade system, the liberalization oftransport activities, and the privatization of state banks and some publicenterprises.

I/ For petroleum products, the domestic sales prices were maintained wellbelow export parity.


- 14 -

2. Liberalization of the exchange system

In The Gambia, at the very start of the floating system, theauthorities eliminated all current account exchange restrictions, \J whilemaintaining export surrender requirements, and the requirement that exchangetransactions be conducted through banking channels. 2/ In fact, theimplementation of the Exchange Control Act was suspended, resultingessentially in the lifting of all restrictions on current, as well ascapital, international transactions. The liberalization of the exchangesystem was completed in July 1990, with the elimination of all outstandingpayment arrears. 3/

In Jamaica, between September 1990 and May 1991, various measures weretaken to ensure that current international transactions were largely free ofexchange restrictions, the remaining restrictions being limited to outflowsof resident-owned capital other than for approved investments abroad. Aspart of these measures, prior approval of the Bank of Jamaica was no longerneeded for bona fide applications for foreign exchange up to certain amountsfor most current invisible payments. The authorities also removed aregulation that required smaller banks to keep their exchange rates close tothat of the two largest banks; this change gave commercial banks greaterscope for setting their rates independently of each other. 4/ Theexchange system was further liberalized in September 1991, with the removalof virtually all restrictions on capital outflows and inflows.

In Guyana, following the official recognition of the parallel so-called"cambio" market (which merged with the new floating interbank market), manytransactions were transferred from the official to the interbank market.However, for a period of about one year the official market continued to besupplied with receipts from traditional merchandise exports, with theproceeds being used for payments on account of public debt service, most

I/ At the same time, as noted later, in relation to trade restrictions,they abolished specific import licensing which was replaced by an opengeneral license (OGL) system.2/ The current account concept used in this paper, derived from balance

of payments methodology, is not identical with the Fund's jurisdictionalconcept of "current international transactions." Although Nigeria andSri Lanka still avail themselves of the transitional arrangements underArticle XIV of the Articles of Agreement, all exchange restrictions oncurrent transactions have been eliminated, at least in the case ofSri Lanka. The Gambia, Guyana and Jamaica have accepted the obligations ofArticle VIII which enjoins members to avoid imposing restrictions on currentinternational transactions without Fund approval.3y Eventually the Exchange Control Act was repealed.4/ As noted earlier, another regulation was removed, whereby forward

exchange rates were derived by using the spot rate as a base plus a specificmarkup.


- 15 -

petroleum products, medical services, some basic food items, and limitedinputs for exporters surrendering in this market. The rates in the twomarkets were unified on February 21, 1991 and the official market wasabolished in September 1991. There remained no restrictions on current andcapital transactions in the interbank market.

The liberalization effort was more gradually phased in the two othercountries surveyed. In Nigeria, with the introduction of the SFEM inSeptember 1986, restrictions on payments for invisibles, which were earliersubject to stringent controls, were eliminated and basic allowances wereprovided for some payments. Most capital account transactions remainedsubject to authorization by the Ministry of Finance. Inward directinvestment was liberalized later in 1988, allowing investors to acquire upto 100 percent in Nigerian ventures, with the exception of banking,insurance and petroleum. Under the SFEM, the commercial banks in Nigeriahad to rely on the central bank administered auctions to obtain access tothe Government's net foreign exchange surplus, namely, the proceeds from oilexports and foreign borrowing minus public sector payments for debt serviceand for imports of goods and other services. I/ Because these largepublic sector transactions were undertaken outside the interbank market, thecoverage of transactions in this market was limited, and the participants inthe market were faced with some measure of uncertainty about the publicsector's planned foreign exchange sales. As the rate in the interbankmarket initially had to be limited to a margin of 1 percent above theprevious auction rate, and as banks tried to bid conservatively at theauction to avoid excessively depreciating the auction rate, 2/ theinterbank market rate was relatively sticky, with the result that therecontinued to be a large informal, parallel market with a more depreciatedrate. The rate in the interbank market (including for the resale ofexchange bought at the auction) was liberalized in January 1987, and itbegan to diverge from the auction rate. From June 1987, banks were requiredto resell the exchange obtained at the auction at the buying price plus a1 percent margin, but a free market-determined interbank rate could beapplied to the funds that banks obtained from autonomous sources. The tworates continued to diverge under this market segmentation and also becauseof new regulations by which commercial banks were required to provide (atthe time of bidding at the auction) information on the expected uses offoreign exchange by their customers, and because they continued to besubject to market share restrictions on the amount of foreign exchange theycould bid for.

JL/ The first-tier administered exchange rate for public transactions waseliminated, as intended, in July 1987, when the rate was unified with theauction rate.

2/ As explained later, conservative bidding by banks in the auctionmarket was largely the result of the application, by the central bank, ofmarket shares in this market that practically guaranteed access to all banksregardless of their bids.


- 16 -

When in 1989 the Nigerian authorities licensed foreign exchangebureaus, without restriction, to compete in the interbank market, theautonomously funded free segment of the interbank market merged with theinformal market, giving rise to a dual exchange rate system--an auction rateand a more depreciated interbank rate in the free (but officiallysanctioned) market that included exchange bureaus. It was understood thatunification of the exchange system would eventually become necessary foreliminating exchange distortions. This was achieved on March 5, 1992 whenbanks and other authorized dealers were set free, without restriction, todetermine the rates in the interbank market and when the Central Bank becameone more supplier of foreign exchange to the interbank market at the goinginterbank rate and ceased relying on limits on foreign exchange allocated tobanks.

In Sri Lanka, the new interbank exchange system was introduced withsome liberalization of the regulations requiring exporters and borrowers torepatriate their external funds through the domestic banks. To promote themobilization of foreign capital, the authorities permitted some domesticentities (such as the Greater Colombo Economic Council and other approvedenterprises) to retain the proceeds of their foreign borrowing in specialaccounts (which were held in Foreign Currency Banking Units). In April1992, a prohibitive tax on the transfer of equity from domestic to foreignownership was eliminated. Increases were authorized in foreign exchangeallowances for business, tourism, and educational purposes; and approval forcapital and dividend repatriation for foreign investors was substantiallyliberalized. In March 1993 the authorities removed restrictions on a rangeof foreign exchange transactions by residents and allowed exporters to keepearnings abroad instead of repatriating them.

In all five countries, although the banks and other authorized exchangedealers were relatively free to negotiate the exchange rate betweenthemselves and the public on a continuous basis, they were neverthelessguided by certain regulations and/or central bank operations that affectedthe price of foreign exchange. As noted earlier, in Sri Lanka regulationspermitted daily exchange rates to vary only within a 3 percent band oneither side of the average rate determined by the market on the previousbusiness day (the reference rate). In Nigeria, the exchange rate in theinterbank market was influenced (until early 1992) by restrictions applyingto that market, as well as by the rate determined at the Central Bankauction (or at the Central Bank "serial" sale after the 1989 reform), overwhich the Central Bank retained considerable control. Except for a briefperiod in 1987, the funds acquired by banks and other foreign exchangedealers at the Central Bank auction or serial sale had to be resold at arate that could not differ by more than 1 percent from the purchase rate.In The Gambia, the Central Bank could influence the rate through itsparticipation (with the banks and other authorized dealers) in the weekly"fixing sessions." In addition, in Guyana, Jamaica, and Sri Lanka, theirrespective central banks reserved for themselves the right to undertakeoperations in the interbank market, including in the latter two cases, forthe explicit purpose of influencing the market rate and not solely for


- 17 -

smoothing out purposes. In Guyana this authority has not been used activelyto date.

The extent to which government or more broadly defined public sectortransactions were conducted at a more appreciated rate relative to thefloating rate in the interbank market differed from country to country. InThe Gambia, the rate for official transactions was determined at weekly"fixing sessions" at which the banks and the Central Bank could buy and sellforeign exchange at freely negotiated rates; the rate emerging from thesefixing sessions was generally close to the prevailing interbank market rate.As Jamaica had introduced a unified floating interbank system, there was nodifferential treatment between public and private sector transactions. InSri Lanka also, the central bank reference rate was a close approximation ofthe interbank "market average" rate, so that public and private sectortransactions were treated in a broadly similar manner. In Nigeria, theadministered first-tier rate which was applied to public sector transactionsuntil July 1987 was generally more appreciated than the interbank marketrate. After unification of the first-tier with the second-tier market,public transactions were effected at the auction rate which was moreappreciated than the free interbank rate (both before and after theintroduction of foreign exchange bureaus). In Guyana, a more appreciatedofficial rate was applied to some public sector transactions such asexternal debt service payments before unification of the rates in February1991.

The extent to which surrender requirements for foreign exchange wereliberalized and the timing of the liberalization varied across the countriessurveyed. In The Gambia, surrender requirements were not formally abolishedbut de facto ceased to be implemented. In Guyana, surrender requirementsfor all exports of goods and services effected through the cambio marketwere eliminated from the start; the remaining surrender requirements fortransactions effected through the official market were eliminated in thecontext of the full unification of the exchange system. In Jamaica, theauthorities eliminated surrender requirements as they removed virtually allcontrols on capital transactions in September 1991. \J In Nigeria, theauthorities eventually allowed non-oil exporters to retain fully theirexport proceeds. In Sri Lanka, most proceeds from exports and borrowingremain subject to surrender requirements to financial institutions; only afew special retention schemes for some public entities were introduced withthe exchange rate float.

The measures that the authorities took to foster competition in theexchange markets differed significantly, and were generally counter-

\J The only remaining capital controls were: (i) prior approval requiredfor the release of funds in excess of J$5,000 a year from blocked accounts;(ii) the regulation that financial institutions match their Jamaica dollarliabilities against Jamaica dollar assets; and (iii) some restrictions oncapital repatriation under the debt-equity conversion program.


- 18 -

productive when they involved regulating the price and/or the volume offoreign exchange transactions. In Jamaica, the Central Bank carried out thefunction of overseeing the working of the market to keep it competitive andefficient in a way that did not directly involve regulating the marketparticipants' activities. By contrast, in Nigeria where the authoritiesbecame highly concerned about the possibility of large banks cornering themarket, they introduced regulations directly designed to prevent thispotential problem. Thus, in connection with the introduction of the 1986exchange system reforms, official limits were imposed on individual banks inthe auction market (5 percent of the total funds offered at each auction forthe three largest banks and 3 percent for the other banks). These limits,in effect, guaranteed widespread distribution of auctioned foreign exchangeamong banks regardless of their bids, which thereby tended to be veryconservative, and proved basically counterproductive in trying to get theauction rate converge to the interbank and free market rates. In the earlystage of the 1989 reforms of the interbank exchange system, the allocationof central bank foreign exchange to banks was made, on the basis of marketshares, \J at the central exchange rate derived from the banks' foreignexchange transactions, but this system was abandoned in December 1990, whena new Dutch auction system was introduced. However, the new system stillretained the restriction that the Central Bank would accept bids from sixseparate categories of banks, with each category having fixed percentageallocations. This arrangement once again was counterproductive and did notsucceed in introducing market forces into the auction market, with theresult that the auction rate continued to diverge from the interbank rate inthe free market comprising the exchange bureaus and the autonomously fundedsegment of the banks. 7J

In Guyana, where explicit competition rules were not developed, cambiomarket participants initially appeared to be wary of the possible politicalimplications of exchange rate depreciation and there was evidence ofinformal rationing by dealers, as there usually was a backlog of foreignexchange demand and as some dealers limited the availability of foreignexchange for particular types of transactions not otherwise regulated. I/Nevertheless, an attempted currency revaluation by market dealers in August1990, against the market fundamentals, did not succeed and was soonabandoned. The existence of a certain backlog of unfilled orders forforeign exchange continues to be a feature of the operations of commercialbank dealers in Guyana. This appears at present to reflect more therelatively undeveloped state of the banking system (no interbank market has

I/ An important reaction to the administrative allocation of foreignexchange to banks after the 1989 reforms was the proliferation of commercialand merchant banks, which benefitted from the allocation of scarce foreignexchange resources at below market prices.2/ As noted above, the unification of the rates was achieved with the

removal of all remaining restrictions in March 1992.3y No evidence of nonprice rationing has been cited in the other four

countries under study.


- 19 -

developed as yet) than a systematic effort to affect the exchange rate,which has exhibited considerable flexibility on several occasions since theintroduction of the cambio market.

Among the five countries surveyed, rudimentary forward foreign exchangemarkets had already operated in Jamaica and Sri Lanka before theintroduction of the interbank market. To foster the forward market, inJanuary 1991 the Jamaican authorities freed it by removing a regulationwhereby forward exchange rates were derived by using the spot rate as a baseplus a specified markup. This regulation, which had been introduced inNovember 1990, had proved to be counterproductive for the development of theforward market. Similarly, in Sri Lanka the direct involvement of theCentral Bank in the forward market, previously available for imports ofessential items, was eliminated; the commercial banks henceforth providedforeign exchange cover for commercial transactions at their own risk and atrates which were freely determined.

3. Reform of the trade regime

In The Gambia, accompanying the liberalization of the exchange ratesystem, the authorities revamped the trade system, putting in place acomprehensive OGL system with no restrictions on product groups orcountries. In 1988 tariffs were lowered in connection with the introductionof a national sales tax of 10 percent which applied to all imports and whichreplaced the previous import tax of 6 percent of the c.i.f. value. In 1989the average ad valorem tariff was further reduced from 28 percent to18 percent. The Gambia Produce Marketing Board's monopoly on the mostimportant export products (in particular, groundnuts) was abolished onFebruary 20, 1990.

In Guyana, on June 15, 1990 (at the same time that additionaltransactions were transferred from the official market to the cambio market)the trade system was simplified and liberalized by removing certain itemsfrom the list of prohibited imports and by removing exemptions from theconsumption tax and import duty previously granted to imports ofagricultural, forestry and mining equipment. Furthermore, import licensesbegan to be granted automatically, within two days of application andwithout restrictions, for bona fide transactions except for imports of fuelwhich required purchases of foreign exchange in the official market.Virtually all import prohibitions were removed at end-October 1990, and inFebruary 1991 Guyana introduced the CARICOM's Common External Tariff. FromJuly 1991 the processing of license applications was further simplified.

Before the introduction of the floating interbank exchange rate system,Jamaica's trade regime had been liberalized in successive steps throughoutthe 1980s, most recently in 1987-88 as part of a trade (and financialsector) reform program supported by the World Bank. Import licenses becamelimited only to foodstuffs, motor vehicles, certain chemicals, fertilizers,pharmaceutical products, wood products, and items endangering public healthor security. All other goods could be freely imported without license.


- 20 -

Concurrently, the structure of trade taxes was simplified so that in generalthe combined tariff and stamp duty applicable did not exceed 10 percent forraw materials, 20 percent for capital goods, and 60 percent for consumergoods. \J On September 1, 1990 an import-duty drawback system replacedthe right of exporters of manufactured goods to receive a rebate of7.5 percent of the value of their exports or the exemption from certainimport taxes.

In Nigeria import licensing was abolished three days after theintroduction of the SFEM in September 1986. However, import (and export)prohibitions were retained for a reduced number of items, including importsof rice, maize, wheat, vegetables, beer, and textile fabrics; andpreshipment inspection was required for most imports valued at US$50,000 ormore.

Paralleling the efforts toward the liberalization of the exchangesystem in Sri Lanka, reforms of the trade regime were undertaken to promoteexports, reduce import protection, liberalize services and allow for greaterscope for capital movements including especially inward direct foreigninvestment. Most state trading monopolies were eliminated, 2J and thelicensing requirements for exports and imports were rationalized andprogressively confined to those maintained for health, environmental, orsecurity reasons. 3/ Tariff reforms aimed at a reduction in thedispersion of effective rates of tariff protection and elimination ofdiscretionary elements in the tariff system. In November 1991, a systemcontaining four bands of 10, 20, 35, and 50 percent was introduced, comparedto a previous system of 19 bands. A new Tariff Commission was appointed in.mid-1992 to make further recommendations on tariffs and other aspects of thetrade system.

IV. Performance of the Interbank Floats

The introduction of floating interbank exchange rate systems and theaccompanying economic reforms in the five sample countries had repercussionson exchange rate and other external sector developments, and on generaleconomic trends. Unfortunately, in the three countries (Guyana, Jamaica,and Sri Lanka) that introduced the new exchange rate system in 1990, therecord for the post-reform years is yet insufficient to speak with fullconfidence on the effects of the new policy upon the external sector andgeneral economic trends. Moreover, in assessing the economic trends, in allcases it is practically impossible to separate the effects of the floating

I/ However, if tariff rates under the Common External Tariff Arrangementof CARICOM (GET) exceeded these amounts, the GET rates would apply.27 The only remaining state trading monopolies are on petroleum and

wheat.3_/ The number of items requiring import licenses was reduced from 143 in

1987 to 93 by mid-1992.


- 21 -

of the exchange rate from other macroeconomic and structural reforms thatwere part of the structural adjustment programs.

1. Exchange rate developments

The behavior of the floating exchange rate varied from country tocountry, as the underlying economic and regulatory conditions varied (Charts1 to 5). I/ The main factors affecting the exchange rate were generaleconomic conditions and, in particular, the rate of domestic inflation;significant modifications of the exchange rate mechanism (when not freelyfloating); and exchange system liberalization measures. In all cases, thepressure of excessive domestic demand which tended to be manifested inrising domestic prices (in varying degrees) also pushed the exchange ratetoward depreciation in the long term.

In The Gambia, Jamaica and Sri Lanka, which from the beginningestablished a unified floating rate and did not significantly alter theexchange rate mechanism after the float, there were continuous short-termchanges in the exchange rate. 2J In The Gambia, after the initialpronounced depreciation, the rate exhibited short cyclical variations withina depreciating trend. In Jamaica, the rate depreciated steadily through theearly months of 1992, when a reversal of the trend took place. In SriLanka, a continuous depreciation similar to that of a de facto crawling pegtook place through nearly the end of 1992.

In Guyana, the rate underwent an initial sharp depreciation with theintroduction of free floating in March 1991, followed by substantialstability until the present.

In Nigeria, the downward slide of the exchange rate was accompanied bythree brief discontinuous depreciations which corresponded to the three mainalterations of the exchange rate mechanism: the initial float in September1986, the revised system in January 1989, and the final unification of theauction and interbank rates in March 1992. There was evidence of somestickiness in the rate in between the discontinuous depreciations, as aresult of constraining regulations described in Section III.

I/ All rates discussed here refer to end-of-month U.S. dollar rates takenfrom International Financial Statistics, except where indicated otherwise.These rates are those reported to the Fund as the main rates in the officialmarket. In Guyana, for the period between the initial float (March 1990)and the open float (February 1991), the rate refers to the fixed officialadministrative rate, not to the interbank rate.

2J For lack of daily data, this refers to the month-to-month changesobserved in Charts I to 5.


- 22 -

The extent of exchange rate volatility under the floating rate regimesvaried across countries. \J The largest degree of volatility, as measuredby the average absolute percentage change in the exchange rate from month tomonth, is encountered in The Gambia 2/ and Jamaica, 3/ which floatedfreely. 4/ By contrast, Sri Lanka's exchange rate exhibited a smallerdegree of variability, partly as a result of the fact that the exchangesystem is a managed float. Guyana's rate has remained remarkably stablesince the free floating in March 1991. In Nigeria, the rate was relativelystable within each stretch in between changes of the exchange ratemechanism.

In The Gambia, Jamaica and Sri Lanka, the floating arrangements andexchange liberalization appeared to have led to the establishment of arelatively efficient foreign exchange market for the allocation of foreignexchange from exporters to importers, in both the public and privatesectors. Access to foreign exchange through official channels was fullyrestored in these countries. In Guyana, by contrast, the relativestickiness of the exchange rate (certainly before the free floating in March1991) was accompanied by some evidence of nonprice rationing by banks; sincethe free float, this rationing appears to be largely related not to anyofficial intervention but to imperfections in the interbank market, such asthe lack of extensive foreign exchange trading among banks and otherforeign exchange dealers. In Nigeria, foreign exchange demands that werenot fully met by banks and other authorized dealers continued to spill overinto the remaining parallel exchange market.

The behavior of the real exchange rates pointed to a real depreciationin all cases but Sri Lanka. The real depreciation was rather large inNigeria, with the real exchange rate index falling to about one fourth theinitial level from 1986 to 1992; this decrease was associated with a sharpdeterioration in the terms of trade resulting largely from the fall in oilprices. In Sri Lanka, the real exchange rate appreciated moderately (froman index of 88 in the third quarter of 1990 to 92 in the second quarter of1992).

I/ The negative effects of such volatility on trade could not bequantified with the available data.

2_/ From April 1986 to September 1992 the average monthly change was2.9 percent. To the extent that The Gambia continued to have a largecomponent of trade in pound sterling, the variability with regard to thedollar may be of secondary importance and may partly reflect the variationsof the pound with respect to the dollar..3/ From October 1990 to September 1992 the average monthly change was

6.6 percent. This includes sharp monthly depreciations in the periodAugust-October 1991 which probably responded to a more accurate estimationof fundamentals by market participants.4/ This abstracts from discrete depreciations resulting from changes in

the exchange rate mechanism in Guyana and Nigeria.


- 22a -

Chart 1: The Gambia: Exchange Rate 1/(Dalasis/US$, 1/83-9/92)

1/ Based on end-of-month data from "International Financial Statistics"


- 22b -

Chart 2: Guyana: Exchange Rate 1/(Guyana$/US$, 1/87-12/92)




Charts: Jamaica: Exchange Rate 1/(Jamaican$/US$, 1/87-9/92)

- 22c -


- 22d -

Chart 4: Nigeria: Exchange Rate 1/(Naira/US$, 1/83-10/92)



- 22e -


Chart 5: Sri Lanka: Exchange Rate 1/(RuDees/US$. 1/87-12/92)


This page intentionally left blank


- 23 -

The effect of the introduction of the interbank exchange rate systemson the parallel exchange rates--and on the breadth of the parallel exchangemarkets--was as expected in all five cases (Table 2). In all cases, theparallel exchange rate premium narrowed substantially by the end of the nextyear after the float, from 33 percent (just before the float) to 3 percentin The Gambia, _!/ from 133 percent (at the end of the year of the float)to 38 percent in Guyana, from 27 percent to 7 percent in Jamaica, from132 percent to 21 percent in Nigeria, and from 24 percent to 1 percent inSri Lanka; those gains were generally maintained or improvedthereafter. 2/

In all five cases the interbank market expanded in terms of the volumeof transactions, as transactions in the parallel market shrank andprogressively became officially recognized. In fact, in all cases butNigeria, exchange reforms have eventually led to the convergence of theparallel and official rates. In The Gambia, Guyana and Jamaica this hasresulted from the elimination of virtually all exchange restrictionsincluding those on capital transactions; surrender requirements no longerexist in Guyana and Jamaica and are not implemented in The Gambia. In SriLanka, the convergence of the parallel and official market rates has beenachieved despite the existence of capital controls and surrenderrequirements for most exports; in this case, relatively tight domesticdemand policies have been a key factor. It is noteworthy that the removal ofcapital controls in The Gambia, Guyana and Jamaica has not undermined thedurability of the floating system. On the contrary, a liberal exchangeregime with free capital movements may have been helpful in preventingprivate destabilizing speculation.

In Nigeria, the continued official restraints on the exchange market,as well as the variability in the restrictiveness of the fiscal and monetarypolicies, were reflected in substantial fluctuations in the parallel marketexchange rate premium; this prompted the authorities to introduce newmeasures--especially the licensing of exchange bureaus to absorb theparallel market--in 1989. These further measures appeared not to have hadthe desired effect on the exchange rate premium, which increased, albeitmoderately, through November 1992; as noted earlier, the modified interbankmarket exchange rate system was not effectively freed from remainingrestrictions, which prevented the adjustment of the official exchangerate. 3_/

I/ Based on the alternative series reported in footnote 1 of Table 2.2/ In Guyana, the premium fell to an insignificant level, according to

new information as indicated in footnote 2 of Table 2.3_/ Moreover, this problem does not appear to have been fully resolved by

the further liberalization of the exchange system and the unification of theauction and interbank rates in March 1992.


- 24 -

Although no systematic data are available on the spreads between thebuying and selling rates following introduction of the interbank exchangerate systems, there is some evidence to suggest that in all cases thespreads either fell considerably or stayed within reasonable limits. In TheGambia, the spread in the interbank market reportedly was around 2 percent,whereas that in the parallel market narrowed in a few months from over8 percent to slightly over 2 percent. In Guyana, the spread in theinterbank market narrowed to within 2 percentage points in a few months. Asimilarly reasonable spread was reported in the interbank market in Jamaica,despite some initial concerns. \J

2. Developments in the external sector and the general economy

In both The Gambia and Nigeria, which introduced the new exchange ratesystem in 1986, external sector data point to an improvement in the currentaccount of the balance of payments (although with a delayed effect in thecase of Nigeria), and an improvement in reserves in relation to imports(Table 3). 2/ The overall balance of payments, by contrast, improved onlyin The Gambia, but not in Nigeria, where substantial private capitaloutflows appeared to have occurred in response to insufficient downwardadjustment of the exchange rate and insufficient upward adjustment of theinterest rate. .3/ All indicators suggest that the modified and moreliberalized interbank exchange rate system introduced in Nigeria in 1989 hada positive effect on external sector adjustment in 1990: a sharp increase inthe current account surplus, a further rise in the ratio of reserves toimports, and an improvement in the overall balance of payments; butpreliminary data for 1991 indicate that there were no lasting positiveeffects, probably because of a lax fiscal policy stance. 4/ In theremaining three countries with more recent floating rate systems,there issome evidence to suggest that the new policies helped to correct the seriousbalance of payments pressures and distortions that had arisen from thepreceding fixed exchange rate system. In Guyana, in particular, the switchto the new liberalized exchange rate regime was followed by a sharpimprovement in the external sector.

I/ Through official action, the authorities in Nigeria limited the spreadin the official market to a 1 percent maximum. Similarly, the spread wasinitially limited to only 0.08 percent in Sri Lanka, but this was raised intwo stages to the more reasonable level of 1 percent by March 1992.

2/ In Nigeria this was largely due, however, to the decline in imports.This in turn reflected the decrease in the value of oil exports as a resultof a substantial reduction in the price of oil, the major export of Nigeria.I/ In The Gambia, relatively large capital outflows in the three-year

pre-liberalization period were followed by moderate inflows or outflows inthe three-year post-liberalization period.4/ This remark also seems to apply to the most recent exchange reform (in

March 1992) which was not complemented by an appropriate tightening offiscal policy. This, together with remaining restrictions, prevented thereduction of the parallel exchange rate premium.


- 25 -

There is also evidence that exchange system reform was not the onlyelement in determining the general performance of these economies;macroeconomic and other structural policies were also important determinantsof the progress toward lowering inflation and increasing economic growth(Table 4). The available data for The Gambia and Nigeria, following theintroduction of their 1986 systems, point to rather different outcomes.There was a substantial reduction in inflation in The Gambia, but nosignificant long-term gains in Nigeria, as a result of its lax domesticfinancial policies. In The Gambia, the broad-based reform effort helpedmaintain the growth momentum, while in Nigeria the rate of economic growthfluctuated widely both before and after the reforms, I/ although itappears to have been strengthened on average after the reforms. In thethree countries that introduced exchange system reforms in 1990--Guyana,Jamaica and Sri Lanka--inflation accelerated initially but, after about oneyear, it tended to decline. In these three cases there are earlyindications of a renewed growth impetus, with good prospects of beingsustainable over the medium term.

V. Conclusions

In all five countries, the period before the introduction of theinterbank exchange rate systems was characterized by a formally peggedexchange rate that was nevertheless periodically depreciated. Their weakeconomic situation reflected--although in various degrees--an unsustainablecurrent account deficit and a virtual exhaustion of official reserves; anaccumulation of arrears, and an intensification of exchange and traderestrictions; substantial inflation and faltering economic growth; andexpansionary financial policies, with generally repressed financial sectors.

In all cases, a market-based floating exchange rate with a liberalizedexchange system was introduced in order to provide a more efficientmechanism for determining the official rate and allocating scarce foreignexchange resources. The interbank market system was chosen over the auctionsystem because it was considered to be more market-based, and because itminimized the scope for administrative interference with the exchange rate.However, a purely free-floating system was initially introduced only in TheGambia and Jamaica, and with a delay in Guyana. To minimize exchange rateinstability, in Sri Lanka free floating was circumscribed within a fixed butadjustable 3 percent band on either side of a reference rate. In Nigeria,the authorities initially introduced a two-tier system comprising a firsttier with an administered rate and a second-tier composite market comprisingboth a central bank auction for commercial banks and a floating interbankexchange market. Because of the regulatory restraints imposed by theauthorities, this system created difficulties for the flexibility of the

\J The major factor was the fluctuation in the production and prices ofoil, not the exchange rate system. However, exchange rate adjustment helpedreduce these fluctuations after the reform.


- 26 -

official and auction rates and for their unification with the interbankrate; there was, in addition, a parallel market rate. In the context of aprocess of liberalization, the official rate was first aligned with theauction rate; eventually the auction and interbank rates were also unified.The introduction of the interbank markets was combined in most cases with asignificant liberalization of the exchange and trade system. Interest ratepolicies were also liberalized either before or soon after the introductionof the interbank exchange system and in most cases this helped to containcapital outflows.

The role of financial institutions in supporting the exchange systemreforms was similar in all countries, but in Nigeria the highly regulatedauction market administered by the Central Bank tended to impede, ratherthan promote, market competition. In all five countries, banks and otherauthorized exchange dealers were generally set free to negotiate theexchange rate between themselves and the public, although this freedom wasconstrained by some exchange regulations. In the case of Nigeria,regulations designed to prevent large financial institutions from corneringthe foreign exchange market had a perverse effect on competition in thismarket. Prudential regulations and market supervision were developed todifferent degrees in the five countries, varying from reporting requirementsto specific limits on foreign exchange working balances and open positions.

The introduction of the floating interbank exchange systems and theaccompanying exchange and trade liberalization facilitated progress toward aunified regime in most cases. The convergence of official and parallelrates was not achieved in Nigeria because it retained some restrictions andbecause a sustained tight stance of financial policies was lacking. Theinterbank rates depreciated in nominal terms in all cases; in real terms anappreciation of the rate occurred only in Sri Lanka and it was small. Theevidence suggests that there was some month-to-month volatility in thefree-floating exchange rates. Finally, while only sketchy data areavailable on the spreads between the buying and selling rates, there areindications that in those countries in which such spreads were unregulated,they stayed within reasonable limits in the interbank market, and that thosein the parallel markets were reduced. This suggests that there is generallyno compelling need to regulate the spreads between buying and selling rates.

The effects of the exchange rate float and exchange liberalization onthe external sector and, even more so, on the general economic performanceare difficult to separate out from those of simultaneous reforms in otherareas and from stabilization efforts. Also, as the interbank exchangesystems were only recently introduced (in 1990) in Guyana, Jamaica, and SriLanka, the record is yet insufficient to examine with full confidence theireffects on economic developments. There are, however, some indications thatthe economic gains from the floating interbank systems accrued more visiblyto the countries that moved early and decisively to abandon previousrestrictions--The Gambia, Guyana, Jamaica and Sri Lanka. By contrast, whenthe efforts were slow and hesitant with regard to exchange liberalizationand macroeconomic stabilization (as in Nigeria), the country continued to


- 27 -

suffer from the distortions of multiple exchange rates and sharpfluctuations in inflation and output.

The main lessons that can be derived from the experience with interbankexchange rate markets in the five countries analyzed are as follows.

(i) Interbank exchange rate markets can operate relatively wellwith a minimum banking infrastructure. As indicated by the experience ofThe Gambia and Guyana, the existence of only a few banks does not constitutean effective impediment to the development of these markets, so long asthere is an effective liberalization of the exchange and trade system toprovide access to exporters and importers. Moreover, the authorities cansuccessfully expand the breadth and scope of these markets by licensingnonbank financial intermediaries and other foreign exchange dealers and byauthorizing them progressively to undertake a full range of foreign exchangetransactions, subject only to prudential limits on foreign exchange workingbalances and open positions.

(ii) These markets can function in an integrated, efficientmanner only if the authorities remove legal and institutional impediments tothe free operation of these markets including, in particular, exchangerestrictions both on current international transactions and on capitaltransfers. As shown by the experience of Nigeria with a composite exchangerate system, segmented exchange markets can continue to exist if regulationsprevent the free flow of resources across market participants. Moreover,any residual official restrictions that may remain will likely give rise tothe continued existence of parallel, informal markets. These parallelmarkets will be the more active, and the rates in them the more divergentfrom the official rates, the more restrictive is the exchange system and themore inadequate the restraint in macroeconomic and financial policies.

(iii) The authorities can foster the development of interbankforeign exchange markets by introducing appropriate prudential regulationand supervision over exchange transactions. Direct regulation of the marketby means of market shares or similar arrangements to promote competition isnot advisable because it tends to be counterproductive for the flexibilityof the exchange rate. A more appropriate policy, as noted above, would bethe licensing of additional foreign exchange dealers.

(iv) In addition to an efficient exchange rate system, a simpleand efficient trade system will also be necessary for the optimal adjustmentof the external sector in order to derive maximum benefits in terms ofeconomic growth.

(v) Finally, the floating of the exchange rate and theliberalization of the exchange system--although helpful in balancing theexternal sector--are not substitutes for the appropriate conduct ofmacroeconomic policies. In the absence of adequate domestic financialrestraint to avoid excess demand pressures and progress toward the free andcompetitive determination of domestic interest rates, exchange rate floating


- 28 -

will lead only to a depreciating exchange rate and continued inflation, withadverse consequences for long-term growth.


Documents

©International Monetary Fund. Not for Redistribution€¦ · assistance from the Fund. The exchange systems of The Gambia, Guyana, Jamaica and Nigeria are currently classified by